A troubled program is far more fixable in month six than in month eleven. Here is what usually goes wrong, what good looks like, and what an independent owner's representative actually does in the first two weeks.
The status deck is still mostly green, but the dates have moved twice and the second slip was explained in language you did not entirely follow. The integrator is professional and responsive, and every question you ask gets a reasonable answer — yet the answers do not add up to a system that goes live when you need it.
Your CFO has told the board a date. Your quality lead has stopped asking about validation because the last three answers were “that comes later.” Your one IT person is now the busiest person in the company. And you have a growing suspicion, which you have not said out loud, that nobody in the weekly meeting is actually representing you.
That suspicion is usually correct. It is also fixable — and the options narrow every month you wait.
Nobody owns the plan on your side. The integrator's project manager works for the integrator. That is not a criticism — it is the org chart. A good one manages their employer's scope, margin, and change orders well, and will do exactly that. Unless you have hired someone whose only obligation is to your outcome, the plan has an author but no owner.
The status report measures activity, not integrity. Green means tasks are moving. It does not mean the requirements still trace, that the decisions made in month two survive the reality discovered in month five, or that the deliverables you have accepted would withstand someone reading them carefully. By the time a red appears, the recovery options have narrowed considerably.
Validation is being deferred into a corner it cannot get out of. “That comes later” is the single most expensive sentence in a regulated implementation. Documentation produced after the fact is too clean, too complete, and too obviously reverse-engineered from the finished system. An inspector can tell. So can a diligence team.
Operations. Does the design still match how the business actually runs — not how it ran when requirements were signed? Most programs drift here first, quietly, and nobody re-checks because the requirements document is treated as settled.
Regulatory. Which parts of this system carry Part 11, Annex 11, or data-integrity obligations, which do not, and is the effort landing in the right places? Over-validating the unregulated parts is not caution — it is budget and calendar taken from the parts that matter.
Capital. Will this system survive a SOX auditor, a diligence process, or a new CFO with a public-company mandate? If an IPO or a private-equity transaction is anywhere in the next three years, controls designed only for today get rebuilt at the worst possible moment.
Global. Is this being built multi-entity and multi-currency, or US-only with a promise to extend later? EU approval after US approval is close to automatic revenue. A system that cannot carry it is not simpler; it is unfinished.
In a troubled program the reported state and the real state have usually separated. So we do not start with a recovery plan.
Days 1–3. Read everything: the SOW and every change order, the requirements, the configuration decisions, the test evidence, the risk log, the last six status reports. Read them against each other. Most of the answer is in the gaps between documents that were each written to be reassuring on their own.
Days 4–7. Talk to the people who are not in the steering meeting — the person doing the testing, the person who will run month-end, the quality lead who stopped asking. They know where the bodies are. They are rarely invited to say so.
Days 8–10. Reconcile the plan against the evidence. Which milestones were actually met against their stated criteria, and which were accepted because the review window lapsed? What is the real remaining scope?
Days 11–14. A written recovery position: what is recoverable, what is not, what the honest date is, what it costs, and which decisions you need to make in the next thirty days — including the possibility that the right answer is to stop, reset the scope, and restart with a different plan.
You get that in writing, with the reasoning visible, whether or not you engage us for anything after it.
If three or more are uncomfortable, the program needs an owner's representative.
Most of these begin with a targeted analysis: one question, answered properly, in writing, inside ten business days. Usually the question is whether the program is recoverable and what it would take. Fixed fee, agreed before we start, and no obligation to do anything with the answer.
If the answer is that you need someone in the owner's seat through go-live, that is our execution oversight line: plan integrity, scope and change-order control, deliverable acceptance, risk governance, validation oversight, and the standing to say a milestone was not met.
We do not implement. Your partner builds; we hold the program to what you actually need. The separation is the point.
We can have an NDA signed the day you ask.